Hepatic decompensation and hepatocellular carcinoma after stopping nucleos (t)ide analogue therapy: Results from a large, global, multi-ethnic cohort of patients with chronic hepatitis B (RETRACT-B study)

Abstract

Background and aims: Despite improvements in management of chronic hepatitis B (CHB), risk of progression to cirrhosis and HCC remains. There is discordance between guidelines regarding the safe discontinuation of nucleos (t)ide analogue (NA) therapy. We aim to analyze rates of liver-related complications (LRC) after NA cessation. Method: Cohort study of virologically suppressed, end-of-therapy HBeAg negative patients with CHB who stopped NAs between 2001 and 2020 from 12 participating centers across North America, Europe and Asia. Survival analysis techniques were used to analyze cumulative incidence of off-treatment LRC (hepatic decompensation and/or HCC) and differences by patient characteristics. LRC was considered related to treatment cessation if diagnosed off-treatment or within 6 months of starting retreatment. Patient was considered cirrhotic at cessation if cirrhosis was diagnosed at any time prior to NA cessation. Results: Among 1528 included patients, 35 developed LRC after NA cessation:18 hepatic decompensation and 17 HCC. Cumulative rate of LRC at 1- and 4-years of off-treatment follow-up was 1.4% and 3.5%. In the LRC group, mean age was 58 ± 11 years, most were male (77%), Asian (91%), Entecavir-treated prior to cessation (66%), infected with HBV genotype B (51%) (genotype C [17%], D [2.9%], A [0%]), and mean HBsAg level at cessation was 2.5 ± 0.7 log10IU/ml. Median offtreatment follow-up was 15 (IQR: 7–37) months. Patients with LRC (vs no LRC) were older at cessation (≥50 years: 80% vs 63%), more often cirrhotic at cessation (37% vs 11%) and had a higher proportion of start-of-therapy (SOT) HBeAg positive disease (29% vs 14%), resulting in higher cumulative incidence of LRC among these groups (Figure). Higher LRC incidence was mainly driven by decompensation for SOT HBeAg positive disease (vs SOT HBeAg negative: adjusted hazard ratio [aHR] 4.3, p = 0.001) while it was driven by both decompensation and HCC for patients aged ≥50 years (vs <50 years: aHR 3.8, p = 0.007) and cirrhotics at cessation (vs noncirrhotics at cessation: aHR 4.7, p < 0.001). There were no significant between-group differences for other patient characteristics. In the LRC group, 4/35 had off-treatment HBsAg loss, and 2/4 had HBsAg loss after the LRC had been diagnosed. Among those without LRC, 109/1493 had off-treatment HBsAg loss. Conclusion: Regardless of off-treatment HBsAg loss, older, cirrhotic, and/or SOT HBeAg positive patients should be carefully assessed prior to stopping NAs to prevent deterioration to LRC.